Gotowa bibliografia na temat „Transfer”

In real-time video processing data transfer between CPU and GPU is a time critical action; time spent transferring data is processing time lost. Several variants of standard transfer methods were developed and evaluated on nine computers and two smart decision algorithms was designed to help choose the fastest method for each occasion. Results showed that the standard transfer methods can be beaten; by using the designed decision algorithms, transfer times between CPU and GPU (both ways) can be reduced by a factor of 7 compared to always using the standard methods.

Asymmetric transfer hydrogenation (ATH) has emerged as a practical, powerful alternative to asymmetric hydrogenation for the production of chiral alcohols, one of the most valuable intermediates in the synthesis of pharmaceuticals, flavour, aroma and agricultural chemicals, and speciality materials. However, the application of ATH has been hampered by its low TONs and TOFs in the past. On the other hand, enabling catalysis in water contributes to one of the most important and challenging fields of modem chemistry, green synthesis. As a solvent for organic reactions, water bears a number of attractive physical-chemical properties over traditional organic molecular solvents. A~ will be seen in the thesis, ATH in neat water has proved to be viable, affording chiral alcohols in fast rates, high productivity and high enantioselectivity. The reduction can be carried out with unmodified or tailor-made catalysts by using mild, readily available formate salt as reductant with no organic solvents required. A series of ligands (L =chiral diamine and its derivatives, amino alcohols) and metal complexes (those of Ru, Rh and Ir) have been explored for the ATH in water, and a wide range of ketones have been efficiently reduced to chiral alcohols (Scheme A-I.) Furthermore, for the PEG-supported catalyst, water as solvent, it can be very easy to separate the catalyst from the product by simple phase separation, which is a very important and ideal dream, especially in pharmaceutical applications. Aldehydes are hard to reduce successfully by TH catalysis in the solvent which is normally used for TH of ketones, such as refluxing 2-propanol, azeotropic formic acid and triethylamine mixture. However, this water-formate reducing system can also apply to the TH of aldehydes, furnishing highly chemoselectivity and remarkably fast reaction rate for the reduction of aldehydes. Thus, the in-water reduction provides a new method that is simple, economical and eco-friendly, and can be readily adapted for laboratory synthesis and commercial-scale production.

Thesis (Ph.D. in Mechanical Engineering)--S.M.U.<br>Title from PDF title page (viewed July 6, 2007). Source: Dissertation Abstracts International, Volume: 67-02, Section: B, page: 1106. Adviser: David A. Willis. Includes bibliographical references.

The purpose of the work described in this thesis was to measure standard electrode potentials and rate constants for various systems, based around cobalt and copper centred catalysts. The intention was to increase understanding of these two different types of complexes, with the knowledge described herein aiding in catalyst and system design. Two main techniques have been used, those of UME voltammetry and SECM. The UME measurements proved technically simple and were used to provide information on redox kinetics and coupled solution reactions. Conversely, the SECM experiments were more complex to perform but allow more complicated systems to be studied. The results obtained in Chapter 3 provide further information about the properties of [Cu] ATP catalysts. Enhanced information could be obtained with better knowledge of the values ofE1I2 or F!., but the work has shown that in organic solvents even simple parameters are difficult to obtain due to complexities such as ion pairing. The theme of Chapter 4; the reaction of [CUll] with Cu o had previously been unexamined. This reaction has been characterised and provides new information of the processes involved in adding Cu o to ATP reaction mixtures. A computer simulation was applied that described the theoretical behaviour of an oxidative etching event. A good fit was obtained between experiment and theory at large values of d/a and at short times. The potentials of a number of [CuI]/[CUIl ]and [Cul]/[CUO] redox couples have been measured in order to understand the substrate potential response during the SECM etching process. The rate of reaction of a number of different [COl] with E2mB has been measured in a range of solvents, as described in Chapter 5. The solvent was seen to have a great effect on the behaviour of each system, stabilising the formation of R[Colll ], such that two reduction processes became evident where mma was the solvent. Such [Co] species are commonly used as C - C bond forming reagents and therefore knowledge of their behaviour in certain solvents is invaluable. Additionally the [Colt E2mB system may be used as an initiator/catalyst system for polymerisation reactions, either in homogeneous or emulsion solutions.

The demands of managing in an international operating environment have changed considerably over the past 20 years due to developments in global markets. Multinational enterprises face fierce international competition and are now tasked not only with developing effective competitive capabilities but also with maximising the knowledge and expertise developed in one part of the organisation by transferring it to another. This work has investigated the position of multinational enterprises in the developing global market and through a broad and thorough review of current literature, identified a gap in the knowledge -a tool for helping the assessment of the transfer of technology prior to the transfer process taking place. Using existing models as a foundation, a new framework has been developed with observations from three case study organisations and the incorporation of other relevant literature. To -make use of the new framework to practitioners it leas been used as the basis of a workbook by which the anticipated difficulties can be judged and a priority focus developed. Validation of the 'r ework has been carried out by a panel of industrialists and academics experienced in international technology transfer. The thesis concludes with a series of recommendations for further work. The originality of this work lies in the development by the author of the new pre-transfer assessment Framework. This should provide clear advantages over previous approaches and give way to improving the success of technology transfer projects.

La demande énergétique de l’humanité augmente rapidement et ne montre aucun signe de ralentissement. Parallèlement à cette problématique, l'utilisation abusive de combustibles fossiles est l'une des principales causes d'augmentation de la concentration de CO₂ dans l'atmosphère. Ces problèmes doivent être résolus en termes de limitation des émissions de CO₂ et de recherche de sources d'énergie renouvelables pour remplacer les combustibles fossiles. De nos jours, l’énergie solaire est l’une des sources d’énergie renouvelables les plus efficaces. La conversion de l'énergie de la lumière solaire en électricité dans le photovoltaïque ou en énergie chimique par le biais de processus photocatalytiques implique invariablement un transfert d'énergie photo-induit et un transfert d'électrons. Dans ce contexte, l'objectif de la thèse est d'étudier les processus photo-induits dans les photosystèmes moléculaires utilisant la photolyse par flash laser. Le premier thème de cette thèse porte sur l’étude du transfert monoélectronique dans des systèmes de dyades donneur-accepteur en vue d’optimiser l’efficacité de la séparation des charges et de son application dans la cellule solaire organique photovoltaïque. Le deuxième thème de cette thèse porte sur l’étude de deux systèmes modèles de photosynthèse artificielle étudiés pour la possibilité d’une accumulation de charge par étapes. Ensuite, différents systèmes photocatalytiques, développés pour la photoréduction du CO₂, ont été étudiés. La compréhension des processus photo-induits devraient permettre l’amélioration de l'efficacité de la réduction du CO₂ dans les systèmes photocatalytiques pratiques<br>The energy demand of humanity is increasing rapidly, and shows no signs of slowing. Alongside this issue, abuse using fossil fuels is one of the main reasons which leads to an increase in atmospheric CO₂ concentration. These problems have to be solved in terms of both limiting CO₂ emission and finding renewable energy sources to replace fossil fuels. Nowadays, solar energy appears as one of the most effective renewable energy sources. Conversion of solar light energy to electricity in photovoltaics or to chemical energy through photocatalytic processes invariably involves photoinduced energy transfer and electron transfer. In this context, the aim of the thesis focuses on studying photoinduced processes in molecular photosystems using laser flash photolysis. The first theme of this thesis focus on studying single electron transfer in Donor-Acceptor Dyad systems towards optimization efficiency of charge separation and application in the photovoltaic organic solar cell. In the second theme of this thesis, two model systems of artificial photosynthesis were investigated to assess the possibility of stepwise charge accumulation on model molecules. A fairly good global yield of approximately 9% for the two charge accumulation on MV²⁺ molecule was achieved. Then, different photocatalytic systems, which have developed for CO₂ reduction, were studied. Understanding of the photoinduced processes is an important step toward improving the efficiency of reduction of CO₂ in practical photocatalytic systems

L‟étude présentée ici tente d‟ouvrir une voie intermédiaire en s‟attachant au processus de transfert d‟innovation entre entrepreneurs sociaux. Nous appuyant sur la littérature sur la stratégie de transfert d‟innovation, qui cherche à développer une perspective globale des stratégies variées utilisées par les entreprises pour transférer des innovations efficacement, nous étudions ce processus de transfert dans un contexte d‟entrepreneuriat social, où l‟équilibre entre objectifs d‟efficience et objectifs sociaux plus larges est explicitement plus grand (Pache et Santos, 2010). Nous mettons ainsi en lumière le fait que certains organismes qui apparaissent et opèrent dans des contextes d‟échec simultané des marchés et des gouvernements axent leurs activités sur la création de valeur. S‟il est vrai que le transfert d‟innovation a le potentiel d‟ajouter de la valeur à toutes les entreprises, les divergences dans les objectifs organisationnels et les identités de celles-ci peuvent aboutir à des processus de transfert très différents, au fur et à mesure que l‟on passe de l‟entrepreneuriat commercial traditionnel à l‟entrepreneuriat social. Les travaux de recherche récents sur l‟entrepreneuriat social se sont centrés sur le besoin qu‟ont ces entrepreneurs d‟augmenter leur impact social en développant d‟avantage leurs opérations réussies (Bloom et Chatterji, 2009), mais le processus de transfert d‟innovations comme stratégie pour augmenter cet impact est mal connu. D‟autre part, la recherche sur le transfert d‟innovation ne s‟est quasiment pas penchée sur cette question, parce- qu‟elle n‟a pas envisagé comment des entrepreneurs qui veulent créer, et non capter, de la valeur s‟engagent dans le processus de transfert. L‟objectif de ce travail est d‟étudier comment les entrepreneurs sociaux gèrent le transfert d‟innovations et s‟organisent pour augmenter l‟impact de leurs innovations. Le sujet ne se limite donc pas aux mécanismes et aux résultats d‟un transfert, comme dans les études précédentes. Les questions posées ici ont pour but de comprendre comment les motivations des entrepreneurs sociaux, que l‟on considère comme différentes de celles des entrepreneurs plus traditionnels, influencent la façon dont se déroule le transfert d‟innovation. Pour ce faire, nous procédons d‟abord au réexamen des objectifs, du contexte, et du périmètre du transfert d‟innovations dans le cas d‟entreprises sociales<br>The purpose of this dissertation is to deepen the study of social entrepreneurship by examining how social entrepreneurs use partnerships and organize themselves to transfer knowledge. My research takes a first step towards building a theory of innovation transfer and scaling for social entrepreneurship. This study builds on three primary foundations. One is the literature on innovation transfer strategy, which seeks to develop a broad view of the various strategies used by firms to transfer innovations effectively. The second is the literature on social entrepreneurship, which highlights the value-creation focus of certain firms that emerge and operate where situations of simultaneous market and government failure exist. Third, I use the literature on institutions, in particular institutional logics, to explain how organizational processes impacted by multiple logics unfold. Using an inductive, multiple-case methodology, I illuminate the innovation transfer process as it relates to social entrepreneurs. My objective is threefold. First, I aim to understand how social entrepreneurs manage the innovation transfer process. Second, I wan to identify the crucial factors influencing innovation transfer between social entrepreneurs, distinguish these factors from those observed in more traditional settings, and examine their impact on the direction of the innovation transfer process. Third, and finally, I hope to offer possible alternative views to the discussion on “scaling” in the social sector, a debate which has, to date, focused on “scaling organizations” as the primary means to expand the scope and reach of innovations developed by social entrepreneurs

Lors de l’apprentissage d’un modèle de classification pour un nouveau domaine cible avec seulement une petite quantité d’échantillons de formation, l’application des algorithmes d’apprentissage automatiques conduit généralement à des classifieurs surdimensionnés avec de mauvaises compétences de généralisation. D’autre part, recueillir un nombre suffisant d’échantillons de formation étiquetés manuellement peut s’avérer très coûteux. Les méthodes de transfert d’apprentissage visent à résoudre ce type de problèmes en transférant des connaissances provenant d’un domaine source associé qui contient beaucoup plus de données pour faciliter la classification dans le domaine cible. Selon les différentes hypothèses sur le domaine cible et le domaine source, l’apprentissage par transfert peut être classé en trois catégories: apprentissage par transfert inductif, apprentissage par transfert transducteur (adaptation du domaine) et apprentissage par transfert non surveillé. Nous nous concentrons sur le premier qui suppose que la tâche cible et la tâche source sont différentes mais liées. Plus précisément, nous supposons que la tâche cible et la tâche source sont des tâches de classification, tandis que les catégories cible et les catégories source sont différentes mais liées. Nous proposons deux méthodes différentes pour aborder ce problème. Dans le premier travail, nous proposons une nouvelle méthode d’apprentissage par transfert discriminatif, à savoir DTL(Discriminative Transfer Learning), combinant une série d’hypothèses faites à la fois par le modèle appris avec les échantillons de cible et les modèles supplémentaires appris avec des échantillons des catégories sources. Plus précisément, nous utilisons le résidu de reconstruction creuse comme discriminant de base et améliore son pouvoir discriminatif en comparant deux résidus d’un dictionnaire positif et d’un dictionnaire négatif. Sur cette base, nous utilisons des similitudes et des dissemblances en choisissant des catégories sources positivement corrélées et négativement corrélées pour former des dictionnaires supplémentaires. Une nouvelle fonction de coût basée sur la statistique de Wilcoxon-Mann-Whitney est proposée pour choisir les dictionnaires supplémentaires avec des données non équilibrées. En outre, deux processus de Boosting parallèles sont appliqués à la fois aux distributions de données positives et négatives pour améliorer encore les performances du classificateur. Sur deux bases de données de classification d’images différentes, la DTL proposée surpasse de manière constante les autres méthodes de l’état de l’art du transfert de connaissances, tout en maintenant un temps d’exécution très efficace. Dans le deuxième travail, nous combinons le pouvoir du transport optimal (OT) et des réseaux de neurones profond (DNN) pour résoudre le problème ITL. Plus précisément, nous proposons une nouvelle méthode pour affiner conjointement un réseau de neurones avec des données source et des données cibles. En ajoutant une fonction de perte du transfert optimal (OT loss) entre les prédictions du classificateur source et cible comme une contrainte sur le classificateur source, le réseau JTLN (Joint Transfer Learning Network) proposé peut effectivement apprendre des connaissances utiles pour la classification cible à partir des données source. En outre, en utilisant différents métriques comme matrice de coût pour la fonction de perte du transfert optimal, JTLN peut intégrer différentes connaissances antérieures sur la relation entre les catégories cibles et les catégories sources. Nous avons effectué des expérimentations avec JTLN basées sur Alexnet sur les jeux de données de classification d’image et les résultats vérifient l’efficacité du JTLN proposé. A notre connaissances, ce JTLN proposé est le premier travail à aborder ITL avec des réseaux de neurones profond (DNN) tout en intégrant des connaissances antérieures sur la relation entre les catégories cible et source<br>When learning a classification model for a new target domain with only a small amount of training samples, brute force application of machine learning algorithms generally leads to over-fitted classifiers with poor generalization skills. On the other hand, collecting a sufficient number of manually labeled training samples may prove very expensive. Transfer Learning methods aim to solve this kind of problems by transferring knowledge from related source domain which has much more data to help classification in the target domain. Depending on different assumptions about target domain and source domain, transfer learning can be further categorized into three categories: Inductive Transfer Learning, Transductive Transfer Learning (Domain Adaptation) and Unsupervised Transfer Learning. We focus on the first one which assumes that the target task and source task are different but related. More specifically, we assume that both target task and source task are classification tasks, while the target categories and source categories are different but related. We propose two different methods to approach this ITL problem. In the first work we propose a new discriminative transfer learning method, namely DTL, combining a series of hypotheses made by both the model learned with target training samples, and the additional models learned with source category samples. Specifically, we use the sparse reconstruction residual as a basic discriminant, and enhance its discriminative power by comparing two residuals from a positive and a negative dictionary. On this basis, we make use of similarities and dissimilarities by choosing both positively correlated and negatively correlated source categories to form additional dictionaries. A new Wilcoxon-Mann-Whitney statistic based cost function is proposed to choose the additional dictionaries with unbalanced training data. Also, two parallel boosting processes are applied to both the positive and negative data distributions to further improve classifier performance. On two different image classification databases, the proposed DTL consistently out performs other state-of-the-art transfer learning methods, while at the same time maintaining very efficient runtime. In the second work we combine the power of Optimal Transport and Deep Neural Networks to tackle the ITL problem. Specifically, we propose a novel method to jointly fine-tune a Deep Neural Network with source data and target data. By adding an Optimal Transport loss (OT loss) between source and target classifier predictions as a constraint on the source classifier, the proposed Joint Transfer Learning Network (JTLN) can effectively learn useful knowledge for target classification from source data. Furthermore, by using different kind of metric as cost matrix for the OT loss, JTLN can incorporate different prior knowledge about the relatedness between target categories and source categories. We carried out experiments with JTLN based on Alexnet on image classification datasets and the results verify the effectiveness of the proposed JTLN in comparison with standard consecutive fine-tuning. To the best of our knowledge, the proposed JTLN is the first work to tackle ITL with Deep Neural Networks while incorporating prior knowledge on relatedness between target and source categories. This Joint Transfer Learning with OT loss is general and can also be applied to other kind of Neural Networks